基于人工免疫網(wǎng)絡(luò)的神經(jīng)網(wǎng)絡(luò)集成方法及其應(yīng)用研究.pdf

1、上海大學(xué)碩士學(xué)位論文基于人工免疫網(wǎng)絡(luò)的神經(jīng)網(wǎng)絡(luò)集成方法及其應(yīng)用研究姓名：張全平申請(qǐng)學(xué)位級(jí)別：碩士專(zhuān)業(yè)：計(jì)算機(jī)應(yīng)用技術(shù)指導(dǎo)教師：吳耿鋒20071219上海大學(xué)碩士學(xué)位論文ABSTRACTThestudyofnetworkensernbleⅡlcoryhaSattractedgreata竹ention鋤dC0nc鋤siIlcemid1990sAlongwitlltllel鶘emllIlberofresearcherSjoillingiIlt

2、llefield，tlleⅡle0DrandapplicationofmeresearchresultshaVepubliShedonebyone，brin百ngmeneuralnetworkeIls翎曲1et0averyaCtivehotspotiIlinterIlatioIlalmacllineleanlingandnemmcomputingItStlleoreticalstudyfocusedoniInpr0Vingmediver

3、sityofens鋤bleAtⅡles鋤etime，artificialimm_unenetwork，弱anewdisciplineofAnificialnellig鉍ce，h嬲developedbyleapsandboundsMallyartificiali加munealgoritllrnswereproduced，觚danumberofapplicationexal玎【plesofarti6cialimmuneneMorkwere∞

4、嘎lblishedT11ispaperc戤矗edoutmerese硼chonartificialinununenetworkbaSedneuralnetworkeIlsembletedmolo翟；y趾ditsapplicationWbproposodaneuraln印7Irork鉍semblemetllodn鋤edAINEN(Anificialhr吼uneNe觚orkbaLsedneuralne帆orkENs鋤ble)basedonan

5、artificialimmunenet、陽(yáng)rk，andexploreditsapplicationonsurfacesubsideIlcepredictionofTbnnelCons仃uctionAtfirst，mepaperintroelucedⅡ1edevelopmaltsituationofrleuralne鉚orkeIlsanbleat110me鋤dabroad，listedⅡlen唧aln鮑vorkensembleapproa

6、Chbycategory，趾df0cllsedonselectivenellralnetwofkensenlble，becauseselectiveensembleistlletlleoreticalbasisofⅡlispaperAtmesametimeweintrodlⅡcedaclassicselectiveinte野ationmethod徹medGASEN，鋤dpoiIltedoutitsi11adequatenessofpau

7、ssiveiIlte霉atcdOnnleomerh觚d，ananificialjmmunea_190ritbm，panicularlyaclaSsicalalgorithmAiNct，isintrodllcedbecausemearti矗cialimmuneIletworl(ismemeoretical10u11dationofmepaperWbpointedoutmeiIladequateIlessofAiNetOneistllatt

8、11ereisn0distiIlctionbetweenBcens觚d鋤tibodies；tlleotheristllat血econverg鋤ceofmutationoperatorisnotrapidenou曲AReriIlnoduCingmetwotlleoriesm饑tionedabove，AINENisputfon)lr蕊whichisanelllalnetworkinte野ationmetllodThepaperintrodu

9、cedmebaSicthinkingoftllememod，mesolutionfor也etwoinadequateIlessofAiNet鋤dtlledetailedstepsofmeal留orithmTwoc0HlpareexperimcIltswereConlpletedbasedontllestall(1aIltdataforAINENandGASENTheeXperiments證cludetlleclassifiedexp舐e

10、nceaIldtllere野essionexperimentTheexperimentresultSslowt11atmegelleralerrorofAINENismuchlowert11anⅡleoneofGASENillre2ressionexp咖ent觚dnleclassifiodexperimentwinlmuchmorecl嬲sesn∞dedt0becalssifiedTllegeneralenorofAINENaIldGA

11、SENisonmes鋤eleVel訪tlleclaSsi6edexp耐mentwiⅡlfewerclaSsesneededtobeclaSsifiedFiIlally，me1、mnelCon鼬mctionsl】rfacesubsidencepredictionquestionisaIlalySedWeeStablishedtllecorrcspondingpredictionmodel，觚dmendidmeeXp甜mentsfollow